The crystal oscillator has been pilfered 😀

Check out some of the WD90C33-ZZ cards here, they all have 14.318MHz crystals on them. The big gold area by the RAM chips is where it's meant to go.
http://www.vgamuseum.info/index.php/cpu/item/ … ital-wd90c33-zz

It looks like this card is configured to use 14MHz from the ISA slot, the crystal appears to never haver been there. Some components necessary to use the crystal are not even fitted.

Ah, I did notice that the solder points looked too similar to the rest of the board. How do you tell it's using the VLB port's clock? Connected pins?

Since that is probably not it, check if the ROM is correct for the card. Reseating just the ROM has helped me get a card working before now.
The gold fingers of the card look like they need a clean too.
Beyond that probably try scorp's method to check for loose pins on the graphics chip: https://www.youtube.com/watch?v=VUx7jZWkYQc

Not on VLB port but on ISA side, the 14MHz clock is connected and the trace is going toward the PLL.

Interesting. I own maybe 50 VLB VGA cards or more, incl. two with WD90C33-ZZ and all of them came with reference 14.318... MHz or similar crystal. Either for dedicated or integrated clock generator.

I thought ISA OSC signal is way to unreliable for this

Can you double check that it is really driven by the bus on this card? The OSC signal is on the solder side of the PCB.

I haven't had time to work on this for a while, but having checked the slot pinout, the osc pin (I assume this is the correct one) is connected on the ISA portion of the slot connector. The crystal does not appear to have been installed in the factory.

Thermalwrong thank you for your suggestions, I'll take a look at that video, and will try cleaning the slot connector.

mpe wrote on 2022-01-05, 22:43:

Interesting. I own maybe 50 VLB VGA cards or more, incl. two with WD90C33-ZZ and all of them came with reference 14.318... MHz or similar crystal. Either for dedicated or integrated clock generator.

I thought ISA OSC signal is way to unreliable for this

Can you double check that it is really driven by the bus on this card? The OSC signal is on the solder side of the PCB.

The ISA OSC signal is perfectly reliable. The reason that some card manufacturers opted not to use it was fear of bus interference noise.

maxtherabbit wrote on 2022-01-12, 15:25:

The ISA OSC signal is perfectly reliable. The reason that some card manufacturers opted not to use it was fear of bus interference noise.

That's kind of contradicting isn't it? It could be either perfectly reliable or noisy, plagued with interference, unreliable at power-on, when shared or even absent on some late ISA motherboards.

Neverhless, it is very rare for a ISA/VLB-era card to not come with the 14.31818 MHz crystal, except perhaps early cards like CGA/EGA. In fact the card from this thread is the first VLB card I've seen without one which I find interesting.

There is a nice blog post about this - http://www.os2museum.com/wp/the-isa-osc-mystery/

>trying to use an oscilloscope as a precision frequency counter
yeah "nice" post

from looking at the photo the ISA OSC pin goes through a via to R58 and from there it should go to, I think from looking at the datasheet, pin 1 of U20 (CH92c68-S) .
Reference frequency for the clock chip being 14.318MHz.